Powers reference wanted

[MartinT]

[Martin Taylor 2014.07.11.10.25]

When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented. But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar. On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

I'm not looking for reasons, of which I can think of at least one, but for a Powers quote or a reference to a quote explaining why QoC should be treated as an intrinsic variable.

Martin

[_Boris_Hartman]

Martin,

MT :
When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented.

HB :
If I understood Bill right, the basics for "reorganization" is Ashby with his "ultrastability" of "essential variables".

MT :
But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar.

HB :
Try in the Bill's book "Making sense of behavior"- The Meaning of Control, 1998. Maybe the Chapter about learning. I hope I understood you right what you want.

MT :
On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

HB :
It's just on the surface. Everything in the organism is sourcing from the same concept.

Boris

···

-----Original Message-----
From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Friday, July 11, 2014 4:32 PM
To: csgnet@lists.illinois.edu
Subject: Powers reference wanted

[Martin Taylor 2014.07.11.10.25]

When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented. But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar. On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

I'm not looking for reasons, of which I can think of at least one, but for a Powers quote or a reference to a quote explaining why QoC should be treated as an intrinsic variable.

Martin

[MartinT]

[Martin Taylor 2014.07.12.07.57]

Martin,

MT :
When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented.

HB :
If I understood Bill right, the basics for "reorganization" is Ashby with his "ultrastability" of "essential variables".

Maybe so, but I was looking for Bill's justification for treating quality of control as one of those essential variables. It's a long time since I read Ashby (maybe 50 years), but I remember being very taken by him. Did Ashby treat quality of control as an essential variable?

MT :
But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar.

HB :
Try in the Bill's book "Making sense of behavior"- The Meaning of Control, 1998. Maybe the Chapter about learning. I hope I understood you right what you want.

I don't have the book, and I am far from home, but if you have a page number, or even better, an actual quote, that would be very helpful.

Martin

···

On 2014/07/12 7:15 AM, Boris Hartman wrote:

-----Original Message-----
From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Friday, July 11, 2014 4:32 PM
To: csgnet@lists.illinois.edu
Subject: Powers reference wanted

[Martin Taylor 2014.07.11.10.25]

When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented. But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar. On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

I'm not looking for reasons, of which I can think of at least one, but for a Powers quote or a reference to a quote explaining why QoC should be treated as an intrinsic variable.

Martin

[_Boris_Hartman]

Martin,

I had a filling that something is not clear what you want, and it occured to me, that it would be good if you explain what you mean by "Quality of Control (QoC).

Boris

···

-----Original Message-----
From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Saturday, July 12, 2014 2:03 PM
To: csgnet@lists.illinois.edu
Subject: Re: Powers reference wanted

[Martin Taylor 2014.07.12.07.57]
On 2014/07/12 7:15 AM, Boris Hartman wrote:

Martin,

MT :
When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented.

HB :
If I understood Bill right, the basics for "reorganization" is Ashby with his "ultrastability" of "essential variables".

Maybe so, but I was looking for Bill's justification for treating quality of control as one of those essential variables. It's a long time since I read Ashby (maybe 50 years), but I remember being very taken by him. Did Ashby treat quality of control as an essential variable?

MT :
But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar.

HB :
Try in the Bill's book "Making sense of behavior"- The Meaning of Control, 1998. Maybe the Chapter about learning. I hope I understood you right what you want.

I don't have the book, and I am far from home, but if you have a page number, or even better, an actual quote, that would be very helpful.

Martin

-----Original Message-----
From: csgnet-request@lists.illinois.edu
[mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Friday, July 11, 2014 4:32 PM
To: csgnet@lists.illinois.edu
Subject: Powers reference wanted

[Martin Taylor 2014.07.11.10.25]

When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented. But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar. On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

I'm not looking for reasons, of which I can think of at least one, but for a Powers quote or a reference to a quote explaining why QoC should be treated as an intrinsic variable.

Martin

[MartinT]

[Martin Taylor 2014.07.12.16.49]

Martin,

I had a filling that something is not clear what you want, and it occured to me, that it would be good if you explain what you mean by "Quality of Control (QoC).

Boris

For a single control unit, Quality of Control can be measured in a variety of ways, but they all come down to the same thing: the ratio between some measure of variation of the controlled perception and the variation that would be observed if only the disturbance influenced the perception. Often the measure of variation used is the RMS variation of the perception. It could be the amplitude. It really doesn't matter in this context.

For a section of the control hierarchy, the QoC could be taken to be the average QoC of the individual elementary control units in that section, or it could be some other combination of the individual QoC values. Again, it doesn't matter in this context which measure you use. When QoC is worse, and especially if it is getting ever worse, the reorganization rate increases. That's the presumption used a lot in Powers's modelling and discussion, and the same is true for any of the intrinsic variables, such as (presumably) body temperature, blood CO2, blood sugar, etc.

All I am looking for is a reference to a place where Powers explains why QoC should act as though it were an intrinsic variable.

Martin

···

On 2014/07/12 9:44 AM, Boris Hartman wrote:

-----Original Message-----
From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Saturday, July 12, 2014 2:03 PM
To: csgnet@lists.illinois.edu
Subject: Re: Powers reference wanted

[Martin Taylor 2014.07.12.07.57]
On 2014/07/12 7:15 AM, Boris Hartman wrote:

Martin,

MT :
When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented.

HB :
If I understood Bill right, the basics for "reorganization" is Ashby with his "ultrastability" of "essential variables".

Maybe so, but I was looking for Bill's justification for treating quality of control as one of those essential variables. It's a long time since I read Ashby (maybe 50 years), but I remember being very taken by him. Did Ashby treat quality of control as an essential variable?

MT :
But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar.

HB :
Try in the Bill's book "Making sense of behavior"- The Meaning of Control, 1998. Maybe the Chapter about learning. I hope I understood you right what you want.

I don't have the book, and I am far from home, but if you have a page number, or even better, an actual quote, that would be very helpful.

Martin

-----Original Message-----
From: csgnet-request@lists.illinois.edu
[mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Friday, July 11, 2014 4:32 PM
To: csgnet@lists.illinois.edu
Subject: Powers reference wanted

[Martin Taylor 2014.07.11.10.25]

When considering reorganization, Powers typically used quality of control (QoC) as an intrinsic variable. Indeed, we had many discussions on and off-line about the detail of how this should be implemented. But I don't know of anywhere he published a justification or reason for why QoC should be an intrinsic variable like, say, body temperature or blood sugar. On the surface, it seems a different kind of concept. So I'm looking for a quote that says something more than that it is a convenient criterion for influencing the rate of reorganization.

I'm not looking for reasons, of which I can think of at least one, but for a Powers quote or a reference to a quote explaining why QoC should be treated as an intrinsic variable.

Martin

[rsmarken]

[From Rick Marken (2014.07.12.1750)]

···

Martin Taylor (2014.07.12.16.49])–

On 2014/07/12 9:44 AM, Boris Hartman wrote:

Martin,

MT: For a single control unit, Quality of Control can be measured in a variety of ways, but they all come down to the same thing: the ratio between some measure of variation of the controlled perception and the variation that would be observed if only the disturbance influenced the perception.

RM: Very close. What you are describing is a “stability” measure of QoC, which is the ratio of observed to expected variance of a controlled variable. The observed variance is, as you say, a measure of variation of the controlled variable: observed var(cv) = var(cv). The expected variance of the cv is the variance of the controlled variable that would be expected if there were no control. In this case the effects of disturbances and control system outputs would have independent (uncorrelated) effects on the controlled variable (when there is control system output are strongly negatively correlated with disturbances). If the state of the controlled variable is known to be determined by the sum of disturbance and output – that is, if cv = o + d – then, since the variance of the sum of uncorrelated variables is the sum of the variances of the variables that are summed, the expected variance of the cv is the sum of the variances of o and d: that is, expected var(cv) = var (o) + var (d). If control is poor, the observed variance of the cv will be nearly the same as the expected variance. So the stability measure of QoC is var(cv)/(var(o)+var(d)).

This approach to measuring QoC only works, however, when the effects of disturbance and output on the cv are known to be additive. In my “What is Size” demo (http://www.mindreadings.com/ControlDemo/Size.html) I computed the stability measure of control of area, a variable where the effects of disturbance and output on the cv were multiplicatve, using logs. If area is the cv then there is evidence of control if the observed variance of the area is much smaller than the expected variance. The observed variance of area is easy to measure; it’s just the variance in area over time. But what is the expected variance of area? In this case the disturbance to area is the height of the rectangle and the output is the width. So area = d * o. Since additivity of variances applies to variables that are a result of adding two variables together it is not true in this case that expected var (area) = var(d) + var(o). However, if we take the log of both sides of the equation for area we get log (area) = log(d) + log(o) so the expected var(log(area)) = var(log(d)) +var(log(o)). The measure of stability I use in the demo is actually S = 1-observed var(area)/expected var(area) so when control is good the ratio of observed to expected variance in area will be very small and S (the stability measure of QoC) will be close to 1.0.

MT: All I am looking for is a reference to a place where Powers explains why QoC should act as though it were an intrinsic variable.

QoC can refer to the quality of controlling any variable, intrinsic or non-intrinsic. But I think what you are referring to is using QoC as a measure of the average error in several non-intrinsic variables in the control hierarchy. I don’t know that Bill ever suggested that QoC in this sense could be an intrinsic controlled variable, but he very well might have. But he does use QoC in this sense as the basis of the reorganization model described in Ch. 7 of LCS III. On p. 117 of that chapter Bill mentions that he is using such as measure as a simplification and that in a real organism the error being minimized by the reorganization system would be “some variable that is important for survival”, ie, an intrinsic variable (like “circulating glucose”). So Bill seems to be ruling out QoC (as you describe it) as an intrinsic variable. But I personally don’t see why QoC (measred as error in the hierarchy) couldn’t be an intrinsic variable that is controlled by the reorganization system. It seems like it is typically this kind of variable – not a “survival” variable – that is the basis of reorganization in psychotherapy.

Best

Rick

–
Richard S. Marken PhD
www.mindreadings.com

[MartinT]

[Martin Taylor 2014.07.13.00.16]

True, but since in the absence of control, var(o) = 0, cv = d.

“Intrinsic controlled variable” sounds like a contradiction in
terms. The intrinsic variables are outside the perceptual control
hierarchy, and are influenced by the side-effects of control. When
they depart from their genetically determined reference values, they
alter the perceptual control hierarchy so that the side effects of
whatever perceptions we happen to be controlling (eventually, but we
hope soon enough) bring them back. That is indeed “control”, but it
is so different from the control in the perceptual control hierarchy
that I think it confusing to use the same word. The control loop
diagram would have an output function that varied its gain all over
the lot, even changing sign on occasion.
The reason I asked whether anyone had a reference for Bill saying
that QoC acts like an intrinsic variable is that I think there is a
very good evolutionary reason that it should do so, and that it
should have done so from as soon in the life story as intrinsic
variables first existed. I wanted to know whether there was a place
I should reference Bill having explained that this is so, and why. Over the last quarter century, I have found all too often that I
thought I had discovered something new about PCT, only to find that
Bill thought of it first. I suspect this might be another such
occasion, so I want chapter and verse to make the point. However,
your comment seems to suggest that Bill actually thought of it as a
convenient simplification. My discussions with him about the
modularity of the measure of QoC as used in reorganization lead me
to think he thought of it as more than that. But I don’t remember
him ever in e-mail saying why.
Martin

···

[From Rick Marken (2014.07.12.1750)]

          Martin

Taylor (2014.07.12.16.49])–

          On

2014/07/12 9:44 AM, Boris Hartman wrote:

Martin,

          MT: For a single control unit, Quality of Control can be

measured in a variety of ways, but they all come down to
the same thing: the ratio between some measure of
variation of the controlled perception and the variation
that would be observed if only the disturbance influenced
the perception.

          RM:  Very close. What you are describing is a

“stability” measure of QoC, which is the ratio of observed
to expected variance of a controlled variable. The
observed variance is, as you say, a measure of variation
of the controlled variable: observed var(cv) =
var(cv). The expected variance of the cv is the variance
of the controlled variable that would be expected if there
were no control. In this case the effects of disturbances
and control system outputs would have independent
(uncorrelated) effects on the controlled variable (when
there is control system output are strongly negatively
correlated with disturbances). If the state of the
controlled variable is known to be determined by the sum
of disturbance and output – that is, if cv = o + d –
then, since the variance of the sum of uncorrelated
variables is the sum of the variances of the
variables that are summed, the expected variance of the cv
is the sum of the variances of o and d: that is, expected
var(cv) = var (o) + var (d). If control is poor, the
observed variance of the cv will be nearly the same as the
expected variance. So the stability measure of QoC is
var(cv)/(var(o)+var(d)).

          MT: All I am looking for is a reference to a place where

Powers explains why QoC should act as though it were an
intrinsic variable.

          QoC can refer to the quality of controlling any

variable, intrinsic or non-intrinsic. But I think what you
are referring to is using QoC as a measure of the average
error in several non-intrinsic variables in the control
hierarchy. I don’t know that Bill ever suggested that QoC
in this sense could be an intrinsic controlled variable,
but he very well might have.

          But he does use QoC in this sense as the basis of the

reorganization model described in Ch. 7 of LCS III. On p.
117 of that chapter Bill mentions that he is using such as
measure as a simplification and that in a real organism
the error being minimized by the reorganization system
would be “some variable that is important for survival”,
ie, an intrinsic variable (like “circulating glucose”). So
Bill seems to be ruling out QoC (as you describe it) as an
intrinsic variable. But I personally don’t see why QoC
(measred as error in the hierarchy) couldn’t be an
intrinsic variable that is controlled by the
reorganization system. It seems like it is typically this
kind of variable – not a “survival” variable – that
is the basis of reorganization in psychotherapy.

[FredNickols]

[From Fred Nickols (2014.07.13.0713 EST)]

This is scary. I actually understood Rick’s notation. Maybe I’m finally learning something.

P.S. Isn’t my date-time stamp a hoot? 0713 0713

Fred Nickols

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Saturday, July 12, 2014 8:48 PM
To: csgnet@lists.illinois.edu
Subject: RE: Powers reference wanted

[From Rick Marken (2014.07.12.1750)]

Martin Taylor (2014.07.12.16.49])–

On 2014/07/12 9:44 AM, Boris Hartman wrote:

Martin,

MT: For a single control unit, Quality of Control can be measured in a variety of ways, but they all come down to the same thing: the ratio between some measure of variation of the controlled perception and the variation that would be observed if only the disturbance influenced the perception.

RM: Very close. What you are describing is a “stability” measure of QoC, which is the ratio of observed to expected variance of a controlled variable. The observed variance is, as you say, a measure of variation of the controlled variable: observed var(cv) = var(cv). The expected variance of the cv is the variance of the controlled variable that would be expected if there were no control. In this case the effects of disturbances and control system outputs would have independent (uncorrelated) effects on the controlled variable (when there is control system output are strongly negatively correlated with disturbances). If the state of the controlled variable is known to be determined by the sum of disturbance and output – that is, if cv = o + d – then, since the variance of the sum of uncorrelated variables is the sum of the variances of the variables that are summed, the expected variance of the cv is the sum of the variances of o and d: that is, expected var(cv) = var (o) + var (d). If control is poor, the observed variance of the cv will be nearly the same as the expected variance. So the stability measure of QoC is var(cv)/(var(o)+var(d)).

This approach to measuring QoC only works, however, when the effects of disturbance and output on the cv are known to be additive. In my “What is Size” demo (http://www.mindreadings.com/ControlDemo/Size.html) I computed the stability measure of control of area, a variable where the effects of disturbance and output on the cv were multiplicatve, using logs. If area is the cv then there is evidence of control if the observed variance of the area is much smaller than the expected variance. The observed variance of area is easy to measure; it’s just the variance in area over time. But what is the expected variance of area? In this case the disturbance to area is the height of the rectangle and the output is the width. So area = d * o. Since additivity of variances applies to variables that are a result of adding two variables together it is not true in this case that expected var (area) = var(d) + var(o). However, if we take the log of both sides of the equation for area we get log (area) = log(d) + log(o) so the expected var(log(area)) = var(log(d)) +var(log(o)). The measure of stability I use in the demo is actually S = 1-observed var(area)/expected var(area) so when control is good the ratio of observed to expected variance in area will be very small and S (the stability measure of QoC) will be close to 1.0.

MT: All I am looking for is a reference to a place where Powers explains why QoC should act as though it were an intrinsic variable.

QoC can refer to the quality of controlling any variable, intrinsic or non-intrinsic. But I think what you are referring to is using QoC as a measure of the average error in several non-intrinsic variables in the control hierarchy. I don’t know that Bill ever suggested that QoC in this sense could be an intrinsic controlled variable, but he very well might have. But he does use QoC in this sense as the basis of the reorganization model described in Ch. 7 of LCS III. On p. 117 of that chapter Bill mentions that he is using such as measure as a simplification and that in a real organism the error being minimized by the reorganization system would be “some variable that is important for survival”, ie, an intrinsic variable (like “circulating glucose”). So Bill seems to be ruling out QoC (as you describe it) as an intrinsic variable. But I personally don’t see why QoC (measred as error in the hierarchy) couldn’t be an intrinsic variable that is controlled by the reorganization system. It seems like it is typically this kind of variable – not a “survival” variable – that is the basis of reorganization in psychotherapy.

Best

Rick

–

Richard S. Marken PhD
www.mindreadings.com

–

Richard S. Marken PhD
www.mindreadings.com

[rsmarken]

[From Rick Marken (2014.07.14.0830)]

Martin Taylor (2014.07.13.00.16)

RM: Very close. What you are describing is a "stability" measure of QoC, which is the ratio of observed to expected variance of a controlled variable. The observed variance is, as you say, a measure of variation of the controlled variable: observed var(cv) = var(cv). The expected variance of the cv is...var (o) + var (d).

MT: True, but since in the absence of control, var(o) = 0, cv = d.

RM: I think you must know that this is not the case. But if not, you can demonstrate to yourself that it's not true using, once again, the "Control of Size" demo (<http://www.mindreadings.com/ControlDemo/Size.html>http://www.mindreadings.com/ControlDemo/Size.html). Suppose that, in this demo, you control area; in that case perimeter is not controlled but var (o) is not zero because the output is being varied to control area. But even though var(o) is not zero, the stability measure for perimeter will be less than that for area because the variance of the perimeter will be approximately equal to the sum of the variances of output and disturbance.

MT: "Intrinsic controlled variable" sounds like a contradiction in terms. The intrinsic variables are outside the perceptual control hierarchy, and are influenced by the side-effects of control. When they depart from their genetically determined reference values, they alter the perceptual control hierarchy so that the side effects of whatever perceptions we happen to be controlling (eventually, but we hope soon enough) bring them back. That is indeed "control", but it is so different from the control in the perceptual control hierarchy that I think it confusing to use the same word. The control loop diagram would have an output function that varied its gain all over the lot, even changing sign on occasion.

RM: It's called "E. coli reorganization, but that is certainly a control process. So I don't see any reason not to say that intrinsic variables are controlled. It's the kind of controlling that you when you control the "dot" in my "Selection of consequences" demo (<http://www.mindreadings.com/ControlDemo/Select.html>http://www.mindreadings.com/ControlDemo/Select.html). You are controlling the movement of the dot even though the results of your outputs are random. This control is implemented by a process (random variation and selective retention) that is, indeed, somewhat different than the control process that is assumed to be used by the perceptual control hierarchy; so it does merit a different name: E. coli reorganization. But the result in both cases is control; keeping a variable in a predetermined state, protected from disturbance. Actually, the resistance to disturbance aspect of "E. coli reorganization" is demonstrated in Marken, R. S. and Powers, W. T. (1989) Random-Walk Chemotaxis: Trial-And-Error as a Control Process. Behavioral Neuroscience, 103, 1348-1355, which is reprinted in my first collection of papers, "Mind Readings".

Which reminds me. All those of you who are interested in the scientific psychology aspect of PCT, don't for get to order copies of my new book, "Doing Research on Purpose". And while you're at it, if you don't already have them, get copies of "Mind Readings" and "More Mind Readings". If nothing else it will be good for the economy (especially mine;-).
Best
Rick

···

--
Richard S. Marken PhD
<http://www.mindreadings.com>www.mindreadings.com

[MartinT]

[Martin Taylor 2014.07.14.12.27]

As usual, we are talking at cross-purposes. I am comparing open-loop

with closed-loop conditions, whereas you are introducing an added
disturbance that is a side-effect of controlling something else.
It’s not surprising we come to different conclusions.
To be talking the same language you should add the effect of varying
the area on the perimeter orthogonally to the externally applied
disturbance, and that produces “d” in the equation.
Exactly my point. Again we use different language to say the same
thing, and appear to be contradicting one another.
E-coli control IS a different kind of control from the control
represented by the usual perceptual control diagram. You can’t specify a gain, and the environmental feedback path may
well keep changing. To me it makes no sense to use the control
diagram to represent e-coli reorganization. All the feedback
influences on the intrinsic variables are caused by side-effects of
perceptual control. The real diagram is rather more complex.
So although we totally agree on what is physically happening, we
manage to seem to be disagreeing. Since this seems to happen a lot,
I wonder if one or both of us might not be controlling for
perceiving a disagreement where none really exists when it is
possible to do so?
Martin

···

[From Rick Marken (2014.07.14.0830)]

            Martin Taylor

(2014.07.13.00.16)

                        RM:  Very close. What you are describing

is a “stability” measure of QoC, which is
the ratio of observed to expected variance
of a controlled variable. The observed
variance is, as you say, a measure of
variation of the controlled variable:
observed var(cv) = var(cv). The expected
variance of the cv is…var (o) + var (d).

            MT: True, but since in the absence of control, var(o) =

0, cv = d.

          RM: I think you must know that this is not the case.

But if not, you can demonstrate to yourself that it’s not
true using, once again, the “Control of Size” demo (http://www.mindreadings.com/ControlDemo/Size.html ).
Suppose that, in this demo, you control area; in that case
perimeter is not controlled but var (o) is not zero
because the output is being varied to control area.

              MT:

“Intrinsic controlled variable” sounds like a
contradiction in terms. The intrinsic variables are
outside the perceptual control hierarchy, and are
influenced by the side-effects of control. When they
depart from their genetically determined reference
values, they alter the perceptual control hierarchy so
that the side effects of whatever perceptions we
happen to be controlling (eventually, but we hope soon
enough) bring them back. That is indeed “control”, but
it is so different from the control in the perceptual
control hierarchy that I think it confusing to use the
same word. The control loop diagram would have an
output function that varied its gain all over the lot,
even changing sign on occasion.

              RM: It's called "E. coli reorganization, but that

is certainly a control process. So I don’t see any
reason not to say that intrinsic variables are
controlled. It’s the kind of controlling that you when
you control the “dot” in my “Selection of
consequences” demo (http://www.mindreadings.com/ControlDemo/Select.html).

[MartinT]

[Martin Taylor 2014.07.14.14.40]

[Martin Taylor 2014.07.14.12.27]

  As usual, we are talking at cross-purposes. I am comparing

open-loop with closed-loop conditions, whereas you are introducing
an added disturbance that is a side-effect of controlling
something else. It’s not surprising we come to different
conclusions.
To be talking the same language you should add the effect of
varying the area on the perimeter orthogonally to the externally
applied disturbance, and that produces “d” in the equation.

I think I didn't word that very well. Let's try again.

A side-effect is an effect on something that is not in the control

loop of the perception being controlled. When you control for
perceiving a glass to be on a table, a side-effect might be its
effect on a perception of the table surface that someone else is
controlling (for example, they might want the table-top to be
empty). Likewise, when in Rick’s demo you control the area, a
side-effect is that the perception of perimeter length may be
influenced. That would be a disturbance to the perimeter perception,
which might also be influenced by whatever external disturbances are
introduce to influence the area perception. The two effects combine
to form the actual disturbance to the perimeter perception (the “d”
in the equation).

Since, by the definition of the problem, the perception of perimeter

is not being controlled, the output of the perimeter control system
is zero, and any changes in the perimeter perception are due to the
joint influences of the disturbance to, and any side-effects of, the
area control system.

···

[From Rick Marken (2014.07.14.0830)]

              Martin Taylor

(2014.07.13.00.16)

                          RM:  Very close. What you are

describing is a “stability” measure of
QoC, which is the ratio of observed to
expected variance of a controlled
variable. The observed variance is, as you
say, a measure of variation of the
controlled variable: observed var(cv) =
var(cv). The expected variance of the
cv is…var (o) + var (d).

              MT: True, but since in the absence of control, var(o)

= 0, cv = d.

            RM: I think you must know that this is not the case.

But if not, you can demonstrate to yourself that it’s
not true using, once again, the “Control of Size” demo (http://www.mindreadings.com/ControlDemo/Size.html

            ).

Suppose that, in this demo, you control area; in that
case perimeter is not controlled but var (o) is not zero
because the output is being varied to control area.